Rheological study of the gel phenomena during epoxide

1

Rheological study of the gel phenomena during epoxide

network formation in presence of sepiolite

Andrés Nohales2, Daniel López3, Mario Culebras1, Clara M Gómez1*

1Institut de Ciencia dels Materials (ICMUV) Universitat de València. Catedrático

José Beltrán, 2. 46980 Paterna, Valencia. SPAIN

2R & D Department, UBE Corporation Europe SA, PO Box 118, 12080

Castellon (Spain)

3Instituto de Ciencia y Tecnología de Polímeros, CSIC. Calle Juan de la Cierva,

3. 28006 Madrid. SPAIN

Con formato: Español (España -alfab. internacional)




2

ABSTRACT

The dynamic behavior during the cross-linking of an epoxy polymer near the gel

point has been monitored by rheological multiple frequency experiments. The

influence of needle-like shape inorganic nanofiller, sepiolite, either non-modified

or organically surface modified during the cure process in presence of an

aliphatic and an aromatic hardener has been investigated. The validity of

different criteria to determine the gel point was examined for the cross-linking of

these filled thermosets. Winter-Chambon criterion at the gel point is obeyed by

the unfilled and by the non-modified sepiolite filled epoxy matrix with either of

the two hardeners. However, physical gels have been formed in presence of the

organically modified sepiolite and the Winter-Chambon criterion is not valid. For

all the systems investigated, the nanofiller reduces the time to reach gelation.

Critical relaxation exponents and gel strength have been determined indicating

more elastic and stronger gel in presence of the aliphatic hardener.

KEYWORDS: epoxy resin, sepiolite, gelation, rheology

3

INTRODUCTION

Epoxy resins are among the best polymeric materials being used in many fields,

especially in the aviation industry as adhesives and in structural applications as

the matrix materials of fiber-reinforced composites. The more outstanding

properties of these matrices are superior strength, stiffness, good chemical and

heat resistance, excellent adhesion and low shrinkage properties. However,

they exhibit low impact strength, poor resistance to crack propagation and small

elongation at break, i.e., they are relatively brittle materials. 1,2

In order to modify epoxy properties, polymer-clay nanocomposites constitute a

new class of materials where nanoscale clay particles are molecularly dispersed

within a polymer matrix 2,3. The interest on these materials stems from the fact

that nanoparticle addition can dramatically change some properties of the

polymer matrix such as tensile strength, heat resistance, gas permeability etc.

with very low loadings of the inorganic filler (1-10 wt%). Much work has been

devoted in obtaining improved matrices with very well dispersed nanoparticles.

However, the factors that control the formation of such hybrids are not well-

understood.

Most of polymer matrix modification with inorganic nanofillers employs a family

of laminar clays, smectite clays, i.e. montmorillonite, MMT. This is a well-known

natural and low-priced layered silicate that shows high aspect ratio, high surface

area, and swelling ability. Because of its high cation exchange capacity easily

turns it organophilic. In order to widen the research of the influence of inorganic

nanofillers on the properties of epoxy matrices, in this work, we investigate the

potential use of sepiolite, a nanofiller of needle-like shape, as potential

4

nanofiller for epoxy resins. Sepiolite is a family of fibrous hydrated magnesium

silicate with the theoretical half unit-cell formula Si12O30Mg8(OH,F)4(OH2)4·8H2O.

4,5 It has a structure similar to the 2:1 layered structure of MMT, formed by two

tetrahedral silica sheets enclosing a central sheet of octahedral magnesia,

except that the layers lack continuous octahedral sheets (Scheme 1). 6,7 The

discontinuous nature of the octahedral sheet allows for the formation of

rectangular channels aligned in the direction of the axis of the particle, which

contain some exchangeable Ca2+ and Mg2+ cations, and “zeolitic water”. These

nanostructured tunnels account in large part for the high specific surface area,

and excellent sorption properties of sepiolite. In addition, it has good

mechanical strength and thermal stability. These properties make sepiolite ideal

for reinforcement of polymer materials, and it has been recently used for the

reinforcement of elastomers 8-11, thermoplastic polymers 12-15, thermosets 16-19

and biopolymers 4.

The cure of a thermoset polymer, such as an epoxy, is a complex process that

involves many chemical reactions. Gelation and vitrification are the significant

phenomena that determine the characteristics of the polymer network. The cure

process begins by formation and expansion of linear chains that soon start to

branch, and subsequently to cross-link, originating three dimensional

networks.1,2 The irreversible transformation from a viscous liquid to an elastic

gel is named “gel point”, that can be defined 1,2,20 as the instant at which the

weight average molecular weight tends to infinity.

5

a)

N+

H2C

CH3b)

Scheme 1. (a) Crystalline and fibre

structure of sepiolite; (b) Surface

organo-modifier of sepiolite B5.

Several theories have been employed to determine the gel point including the

determination of divergence point of the steady state shear viscosity (η) or

normal stress (N1), the cross-over point of G’ (dynamic elastic modulus) and G”

(dynamic viscous modulus) 21. However, Winter-Chambon criterion is commonly

used in polymer gels. In accordance with Winter-Chambon criterion, the loss

tangent (tanδ) is independent of the frequencies at the gel point for most of the

polymers. The frequency-independence of the loss tangent in the vicinity of the

gel point has been widely used to examine chemical and physical gels and has

also been employed to determine the gel point. 21-24 The question is whether or

not the Winter-Chambon criterion would be applicable for the cross-linking of

epoxy thermosetting polymers in the presence of sepiolite.

The aim of this study concerns the measurement of the dynamic viscoelastic

properties of epoxy-amine networks buildup as a function of curing time and

oscillatory frequency. Different methods employed to determine the gel point of

6

an epoxy resin have been tested when these systems are modified by an

inorganic nanofiller, sepiolite. Two different amine curing agents, one aliphatic

and the other aromatic, have been used to cure the epoxy resin.

Consequently, monitoring the rheological behavior of sepiolite-filled epoxy

resins during the cure reaction may lead to better control of solid state

properties. Therefore, in this work, we varied two parameters: 1) the type of

sepiolite used, non-modified and surface-functionalized, and 2) the curing

agent, aliphatic or aromatic, since the choice of curing agent and conditions

would determine the final material properties.

7

EXPERIMENTAL

Materials

The epoxy resin used was Araldite® GY250, a commercial diglycidyl ether of

bisphenol-A kindly provided by Vantico Spain. The epoxy content was 5.34

eq/kg, as determined by acid titration, with a weight per equivalent of 187.3 g/eq

and a hydroxyl/epoxy oligomer molecules ratio of 0.122. The curing agents

were a diamine terminated polypropylene oxide Jeffamine D230 (D230) kindly

supplied by Huntsman Corporation (equivalent [H] weight: 57.5) and an

aromatic diamine 4,4’-methylene-bis-(2,6 diethyl) aniline (MDEA) supplied by

Lonza Cure. Sepiolite Pangel® kindly supplied by Tolsa (Spain) was used as

nanofiller. The dimensions of a single sepiolite fibre vary between 0.2 and 3 μm

in length, 10-30 nm in width and 5-10 nm in thickness, with an average aspect

ratio of about 27. Sepiolite has a surface area of about 300 m2g-1, and a surface

energy of about 240 mJm-2 . 25 The clay modifiers were the unmodified sepiolite

Pangel S9 and the organically modified sepiolite Pangel B5 (a surface modified

sepiolite with benzyl dimethyl alkyl ammonium in order to make it more

compatible with low polarity polymers). (see Scheme 1)

Sample preparation

Sepiolite dispersions in the epoxy pre-polymer were obtained as follows: the

clay dried in an oven at 120 °C was slowly added to the required amount of

epoxy pre-polymer and then dispersed by a high-speed mixer Dispermat® CN20

(VMA-Getzmann GMBH, Germany) at 7200 rpm with a 5 cm diameter disk for

25 min. The mixture was completely degassed under vacuum at 75 ºC. The

Eliminado: clay

8

sepiolite loading expressed as grams of sepiolite per hundred grams of resin

(phr) were as follows: 0, 2.5, 5.0 and 7.5 phr. The curing agent (Jeffamine D230

or MDEA) was added to the sepiolite/pre-polymer dispersions at a

stoichiometric ratio epoxy/amine, [E]/[H]=1, then the mixture was stirred for 8

min, degassed and poured out between the rheometer plates to carry out the

rheological experiments.

Rheological Measurements

Rheological oscillatory measurements were performed on a controlled stress

rheometer TA Instruments AR1000N, using the parallel plate (25 mm diameter)

shear mode to measure the storage modulus, G’, the loss modulus, G’’ and the

loss tangent, tan δ. The thickness of the samples was about 1.0 mm. In order to

reinitialize the dispersions, a steady pre-shear was performed, at a shear rate

of 500 s-1 for 5 min, at room temperature, followed by a 120 min resting time,

before each dynamic experiment.

The rheological study of the curing process was performed at 75 ºC for the

Jeffamine D230 system and at 130 ºC for the MDEA system. Time sweeps

were carried out in a multi-frequency experiment (frequencies between 5 and 50

rad/s). The applied oscillatory torque was varied between 10 and 100 μNm at

increasing curing times to assure that the experiments were performed within

the linear viscoelastic region.

Eliminado: Before each dynamic experiment,

Eliminado:

9

RESULTS AND DISCUSSION

Multifrequency rheological experiments were performed during the buildup of an

epoxy network cured with either an aliphatic or an aromatic diamine in absence

and presence of sepiolite to determine rheological behavior near the gel point.

Several models have been proposed in the literature to study gelation. Among

them, three have been selected in this paper. These are:

1) based on scaling laws, the gel point is the point of crossover of the elastic,

G’, and the loss, G”, moduli in small-amplitude oscillatory experiments 26;

2) Winter showed 20, 23, 27, 28 that the first definition was not universal but only

valid for step-growth polymerizations of balanced stoichiometry. Accordingly, a

more general method of detecting the gel point from dynamic moduli data is

based on the observation 23 that, at the gel point, tanδ is independent of

frequency and is given by:

⎟⎠⎞

⎜⎝⎛ π

=ωω

=ωδ2

ntan)("G)('G)(tan (1)

where n is the relaxation exponent and can be linked to structural parameters of

the network as we shall explain later. The value of tanδ independent of

frequency enables us to determine the value of n. In the special case where

n=1/2, it follows from eq 1 that tanδ=1, G’ crosses over G” and the gel point can

be determined from G’ and G” plots as in model 1.

3) In addition, and based on percolation models 29 in the region close to the gel

point the elastic and loss moduli can be represented by the power laws:

Eliminado: of

10

'n'GA)('G ω=ω (2)

"n"GA)("G ω=ω (3)

being AG’ and AG” the pre-exponential factors, ω the angular frequency and n’

and n” the power law exponents. The gel point is attained when n’=n” = n.

The relaxation exponent can be related to structural parameters of the network

such as the fractal dimension, df and the gel strength, Sg. Thus, Muthukumar

proposed the following expression to relate the exponent n with excluded-

volume and hydrodynamic screening effects 30:

)d2d(2)d22d(dn

f

f−+

−+= (4)

where d=3 is the Euclidean space dimension. All values of the relaxation

exponent for 0<n<1 are possible for a fractal dimension in the physically

acceptable range 1≤df ≤ 3.

Winter and coworkers20,23,28 can describe the rheological behavior at the point

of gelation in terms of the gel strength of the material at the condition of critical

gelation:

(5)

The shear relaxation G(t) modulus is expected to obey a power law relaxation at

the gel point :

ngtS)t(G −= (5)

11

where t is the relaxation time and the equation assumes that there is no

chemical reaction occurring. The parameter n is the critical relaxation exponent

and determines the stress relaxation rate at the gel point. Sg is the gel strength

and can be simply considered as the relaxation modulus at the gel point when

the relaxation time is one second. Sg represents the elastic modulus or the

rigidity of the system when n=0, while it represents the viscosity when n=1. Sg

can be calculated from equation 5 at the gel point, if n is known.

10-1

101

103

105

G´ (

Pa)

G

´´ (P

a) a)

increasing ω

b)

increasing ω

90 92 94 96 98 100 102 104 106 108 110

d)

time (min)

increasing ω

44 48 52 56 60 6410-1

100

101

102

c)

tan

(δ)

time (min)

increasing ω

Figure 1. Logarithmic plot of the storage G’(ω), loss, G”(ω), moduli (parts a and

b) and tanδ (parts c and d) for isothermal curing as a function of time for

different angular frequencies from 5 to 50 rad s-1. The systems studied are

epoxy/D230 cured at 75 ºC in parts a) and c), and epoxy/MDEA cured at 130 ºC

in parts b) and d).

12

Rheological behavior of the epoxy/amine systems during gelation

Gel time on the epoxy systems was investigated at constant temperature

through multi-frequency oscillatory experiments. Figures 1a and 1b show the

evolution of the storage modulus, G’, and the loss modulus, G’’, as a function of

time for different oscillatory frequencies (5-50 rad/s) for the epoxy pre-polymer

cured with Jeffamine D230 at 75 ºC (Figure 1a) and with MDEA at 130 ºC

(Figure 1b). It is observed that the gelation time defined by the time of crossover

of G’ and G”, increases as the oscillatory frequency of the experiment

increases. Nevertheless, the gelation time is an intrinsic property of the system

and consequently cannot be dependent on the experimental conditions 1,2,31. In

order to determine the gel point another criterion has to be employed. Figures

1c and 1d show the evolution of tanδ as a function of time for different

oscillatory frequencies (5-50 rad/s) for the systems epoxy/Jeffamine D-230 at

75 ºC (Figure 1c) and for the system epoxy/MDEA at 130 ºC (Figure 1d). As

expected, values of tanδ increase as frequency increases before the crossover

point, and decrease after it. As can be observed, the value of tan δ diminishes

with time for all the frequencies studied. However, the decreasing rate of tan δ

is frequency dependent, being faster for the lower frequencies. At a certain

point all the curves intersect, that is, tan δ becomes frequency independent.

Before the crossover point, tan δ decreases as the frequency increases, and

after the crossover point tan δ increases with frequency. Changes in the

material structure have been attained. According to the Winter and Chambon

method 23, the gel time can be defined as the time for tan δ to be frequency

Comentario [UdW1]: Both G’ and G” increase as frequency increases before the gel point, with G” values higher than G’ ones. After the gel point G’ values are higher than G” values indicating a more rigid structure. ¿Se puede poner algo asi? Yo pondría a solidlike structure en lugar de a more rigid structure

Eliminado: usually

Comentario [UdW2]: ¿Se puede poner algo asi? Puedes escribir algo mas certero. Es correcto y se puede poner la referencia de Nijenhuis (31)….aunque es repetitivo porque luego se describe mejor en el texto

13

independent. The application of this method provides gel points of 52.6 min and

101.1 min for the epoxy/ D230 and epoxy/MDEA systems, respectively. These

values are in good accord with published results 32,33.That is, reported gel

values determined as the time for G’=G” are 57 min when curing at 80 °C with

polyoxypropylenediamine (Jeffamine D400) 32 and 115 min when curing at 80

°C with MDEA33.

Owing to the structural changes taking place during the gelation process, a

study of their evolution was carried out in the vicinity of the gelation point. The

rheological behavior of the systems as a function of frequency and the evolution

of the critical exponents as a function of time, and its relative distance to the gel

10 100101

102

103

104

105

G´ (

Pa)

ω (rad/s)

a)

increasing curing time

10 100101

102

103

104

105

G´´

(Pa)

ω (rad/s)

b)


10 10010-1

100

101

102

103

104

105

G´(

Pa)

ω (rad/s)


c)

10 10010-1

100

101

102

103

104

105

G´´

(Pa)

ω (rad/s)


d)

Comentario [UdW3]: Revisa el comentario

14

Figure 2. Double logarithmic plot of G’ and G” obtained by interpolation of

experimental data at different times around the gel point as a function of

frequencies. The systems studied are epoxy/D230 cured at 75 ºC in parts a)

and b), and epoxy/MDEA cured at 130 ºC in parts c) and d).

point were studied. One of the problems of this kind of study comes from the

fact that the points corresponding to each frequency are obtained at increasing

times and, therefore, the value of the rheological variables are acquired at

different degrees of connectivity as the cross-linking reaction progressed with

time. To overcome this problem, it was necessary to interpolate the

experimental data to obtain values of G’ and G’’ at the same time, and therefore

to study the behavior of the system as a function of frequency. The interpolation

was carried out through the fitting of the experimental results to an eighth-order

polynomial 34. Figure 2 depicts the results obtained from the interpolation of the

rheological data as the double logarithmic plot of G’ and G’’ as a function of

frequency at different reaction times in the vicinity of the gelation point, for the

system epoxy/ D230 (Figures 2a and 2b); and for the system epoxy/MDEA

(Figures 2c and 2d). The following observations can be drawn from these

figures: i) the double logarithmic plots show a linear dependence of G’ and G’’

with frequency, both G’ and G’’ increase with frequency, ii) the values of G’ and

G’’ increase as the reaction time increases, and iii) the slopes of the variation of

G’ and G’’ with frequency, defined as n’ and n’’, respectively, decrease with the

reaction time and become equal at the gel point. These changes show the

switching from a liquid state, characterized by the dependence of the elastic

15

modulus with frequency, to a solid state, characterized by the independence of

the elastic modulus with frequency. The use of the criterion: G’ ∝ G’’ ∝ ωn at the

gel point 29 gives values of gel times (tgel) of 52.3 min and 101.0 min and of n=

0.659 and 0.628 for the systems epoxy/ D230 32 and epoxy/MDEA 33,

respectively.

The application of equation 4 allows30 the determination of the fractal dimension

of the networks formed and provides values of 1.80 and 1.84 for the

epoxy/D230 and epoxy/MDEA systems, respectively. These values are in

agreement with the occurrence of entropic networks made up of cross-linked

flexible chains. Moreover, following equation 5 values of Sg are 522 and 360 for

epoxy/Jeffamine D230 and epoxy/MDEA, respectively. As Sg depends on chain

flexibility and cross-linking density 23, this result suggests stronger gels for the

Jeffamine system probably due to more stiffen chains and higher crosslinking

densities.

The effect of the type of hardener on the properties of the cured epoxy systems

can be resumed as follows: networks obtained with the amine D230 are formed

at shorter times, have higher gel strength values and are somewhat more rigid

than networks obtained with the amine MDEA.

Rheological behavior of the unmodified S9 sepiolite / epoxy /amine

systems during gelation

The effect of the unmodified sepiolite, S9, on the rheological behavior of the

system epoxy/amine during gelation was investigated. In Figure 3, the evolution

Comentario [DL4]: Yo no diría que son más rígidos, de hecho no se mide en ningún momento la rigidez de los materiales. Yo hablaría de que presentan un mayor módulo o son más elásticos….como de hecho decís en las conclusiones.

16

of tan δ as a function of time, at different oscillatory frequencies (5-50 rad/s), for

the system epoxy/ D230/ sepiolite-S9 at 75 ºC (Figure 3a) and for the system

epoxy/MDEA/ sepiolite-S9 at 130 ºC (Figure 3b) are represented. The sepiolite

concentrations in both systems were 0.0, 2.5, 5.0 and 7.5 phr. The results

shown in Figure 3 indicate that the rheological behavior of the epoxy system is

affected by the presence of sepiolite. On the one hand, the gel time determined

according to the criterion of Winter and Chambon (tanδ independent of

frequency) 23, decreases with sepiolite concentration and, on the other hand,

the value of tan δ at the gel point does also depend on the clay concentration.

The values of the gel point and the critical exponents for the epoxy /amine/

sepiolite S9 systems are collected in Table 1. Gel times determined by the first

criteria, G’= G”, are slightly higher than the ones determined by the two other

criteria. Either Winter-Chambon or percolation models can be used to determine

gelation in unmodified epoxy/ amine/ unmodified sepiolite.

30 35 40 45 50 55 600.1

1

10

100

tan

(δ)

time (min)

a)

0 phr

2.5 phr

5 phr

7.5 phr

60 70 80 90 100 110 1200.01

0.1

1

10

100

time (min)

0 phr2.5 phr

5 phr7.5 phr

b)

tan

(δ)

Figure 3. Evolutions of tanδ obtained at different frequencies and sepiolite S9

loadings (in phr) for: a) epoxy/D230 cured at 75 ºC, and b) epoxy/MDEA cured

at 130 ºC.

Con formato: Inglés (EstadosUnidos)

17

Whatever the hardener used, the gelation time decreases with the

concentration of sepiolite and is more accentuated as the concentration of

sepiolite increases. The decrease in gelation time may be due to both: i) an

increase in the reaction rate due to a catalytic effect of the sepiolite on the

cross-linking reaction, or ii) a decrease in the critical conversion of gelation due

to the presence of the particles of sepiolite organized as a physical network17,35-

37.

To study how the structural changes in the physical network of sepiolite

particles can affect the gelation process, a study of the evolution of the critical

exponents in the vicinity of the gelation point was carried out. As for the pure

epoxy/hardener systems presented above, in Figure 4 the double logarithmic

plot of G’ and G’’ as a function of frequency is shown for the systems sepiolite

S9/ epoxy (concentration 5 phr, as an example) cured with Jeffamine (Figs. 4 a

and b), and with MDEA (Figs. 4 c and d). As can be observed, both G’ and G’’

moduli increase with curing time and show a linear dependence with frequency

in the double logarithmic plot. This behavior, similar to the pure epoxy/hardener

system, indicates an exponential dependence of the moduli with frequency. In

Figure 5, the evolution of the critical exponents n’ and n’’ as a function of the

relative distance to the gelation point, ε, are represented for the systems

epoxy/amine/ S9 sepiolite at different concentrations of the filler (Figures 5 a

and b for Jeffamine D-230 and Figure 5 c and d for MDEA). ε is a dimensionless

parameter characterizing the region around the gel point defined as: ε=|tg-t|/t;

where t is any time during the reaction; and tg is the gel point time. Initially, well

below the gelation point, the viscoelastic behavior of the systems without

18

sepiolite is typical of Newtonian liquids with n’=2 and n’’=1. As the reaction

proceeds, n’ and n’’ diminish and become equal at the gelation point. After that,

the system tends to a hookian behavior with n’=0 and n’’=1. 38

The addition of sepiolite affects the values of n’ and n’’ that diminish with the

concentration of sepiolite for both the Jeffamine and the MDEA systems. This

behavior is probably due to interactions between the sepiolite particles and the

pre-polymer of the type hydrogen bonding or hydrophobic interactions 39. These

interactions provoke the sepiolite particles to be interconnected through weak

links giving rise to a physical network and therefore to an increase of the elastic

101

102

103

104

105

106


G´ (

Pa)

a)

101

102

103

104

105

106

G" (

Pa)


b)

10 100101

102

103

104

105

106


G´(

Pa)

ω (rad/s)

c)

10 100101

102

103

104

105

106

G" (

Pa)


ω (rad/s)

d)

Figure 4. Double logarithmic plot of G’ and G” obtained by interpolation of

experimental data at different times around the gel point as a function of

19

frequencies. The systems studied are epoxy/D230/ 5phr S9 cured at 75 ºC in

parts a) and b), and epoxy/MDEA/ 5phr S9 cured at 130 ºC in parts c) and d).

modulus and to a quasi solid-like behavior. If we consider the evolution of the

critical exponents n’ and n’’ with the reaction time, it can be observed that: well

below the gelation time, n’ for the epoxy/amine/S9 sepiolite systems increase

with the reaction time, unlike for the pure epoxy/amine system. After that, n’

reaches a maximum and finally diminish monotonously till becomes equal to n’’

at the gelation point. The increment of n’ with time indicates a departure from

the solid-like behavior to a more viscous behavior. This can be explained

considering a breakage of the physical network of the sepiolite particles as the

cross-linking reaction of the epoxy system proceeds. A competition between the

rupture of the physical network of the filler particles and the chemical network

formation of the epoxy/amine system is established that would account for the

evolution of the critical exponents with reaction time.

20

0.0

0.5

1.0

1.5

2.0 0.0 phr 2.5 phr 5.0 phr 7.5 phr

Crit

ical

exp

onen

t, n´

a) 0.0 phr 2.5 phr 5.0 phr 7.5 phr

b)

-0.2 -0.1 0.0 0.1 0.20.0

0.5

1.0

1.5

2.0

criti

cal e

xpon

ent,

n´

ε

0.0 phr 2.5 phr 5.0 phr 7.5 phr

c)

-0.2 -0.1 0.0 0.1 0.2

ε

0.0 phr 2.5 phr 5.0 phr 7.5 phr

d)

Figure 5. Variation of the power law exponents, n’ and n”, as a function of ε for

epoxy/D230/ S9 cured at 75 ºC in parts a) and b), and epoxy/MDEA/ S9 cured

at 130 ºC in parts c) and d).

Both, the increase of n’ with the reaction time and its decrease with the sepiolite

concentration indicate that the sepiolite/epoxy/amine system is consistent with a

physical network of sepiolite nanoparticles that evolves to a chemical network of

the epoxy polymer with dispersed nanofiller as the cross-linking reaction takes

place.

The characteristic parameters of the gelation point, that is, the exponent n, the

corresponding tanδ, and the gel strength, Sg, are also dependent on the

sepiolite concentration as can be observed in Table 1. n values are similar to

21

the ones determined by the percolation theory, n= 0,7 and the Rouse model,

n=2/3. 34,39 An important fact is that the evolution of Sg and n with sepiolite S9

concentration is different depending on the hardener considered. For the

system epoxy/D230/ S9, the relaxation exponent n diminish, and the gel

strength Sg increases with filler concentration indicating a more rigid and elastic

gel and therefore, a more compact structure. According to the Muthukumar 30

equation a decrease in the relaxation exponent implies an increase in the fractal

dimension of the gel. On the other hand, for the system epoxy/MDEA/S9 the

behavior is the opposite: the relaxation exponent n increases, the fractal

dimension decreases, and the gel strength decreases with filler concentration

indicating a less rigid, less elastic and more open structure.30, 40 The cause for

the different behavior can be found at the interaction level being the Jeffamine

D230 hardener more polar than MDEA and consequently giving rise to stronger

interactions with the sepiolite filler. 41

Rheological behavior of the modified B5 sepiolite / epoxy /amine system

during gelation

A similar study was carried out with an organically surface modified sepiolite

(sepiolite B5) with benzyl dimethyl alkyl ammonium. In Figure 6, the evolution of

tan δ as a function of time, at different oscillatory frequencies (5-50 rad/s) for the

system epoxy/ D230/ sepiolite-B5 at 75 ºC (Figure 6a) and for the system

/epoxy/MDEA/ sepiolite-B5 at 130 ºC (Figure 6b) are represented. The sepiolite

22

concentrations in both systems were 0.0, 2.5, 5.0 and 7.5 phr. The observation

of these Figures allows us to extract important differences in relation to the

systems with unmodified sepiolite: i) the value of tan δ diminish with frequency

for both hardeners at any filler concentration; ii) the values of tan δ increase with

reaction time, goes through a maximum and then diminish; iii) the decrease of

tan δ is frequency dependent, being faster for the low frequencies, nevertheless

contrary to the systems with unmodified sepiolite, the curves corresponding to

different frequencies do not intersect, that is, tan δ do not become frequency

independent, except for the sample epoxy/D230/ B5(2.5phr). The last effect

becomes more evident as the concentration of sepiolite B5 increases.

Experiments of G’ and G” of epoxy-sepiolite dispersions as a function of

frequency show a higher G’ than G” value at concentrations higher than 5 phr in

presence of organically modified sepiolite. Moreover, FTIR determinations (not

shown here) indicate no chemical cross-linking prior to the rheological

experiments. All these statements point to the formation of a physical network of

sepiolite particles previous to the curing process, being the transition from a

physical to a chemical network more apparent than for the unmodified sepiolite

system.39

23

20 25 30 35 40 45 50 55 600.1

1

10

100

tan

(δ)

time (min)

a)0 phr

2.5 phr

5.0 phr

7.5 phr

60 70 80 90 100 110 1200.01

0.1

1

10

100

tan

(δ)

time (min)

b)

0 phr

2.5 phr

5.0 phr

7.5 phr

Figure 6. Evolutions of tanδ obtained at different frequencies and sepiolite B5

loadings (in phr) for: a) epoxy/D230 cured at 75 ºC, and b) epoxy/MDEA cured

at 130 ºC.

In Figure 7, the evolution of the critical exponents n’ and n’’ as a function of the

relative distance to the gelation point are represented for the systems

epoxy/amine/ sepiolite B5 at different concentrations of the filler (Figures 7a and

7b for Jeffamine D-230 and Figure 7c and 7d for MDEA). As can be observed,

the decrease of the critical exponents n’ and n’’ in relation to the pure

epoxy/amine systems is more dramatic with the modified B5 than with the

unmodified S9 sepiolite. Owing to the surface modification of the sepiolite, the

interactions between the pre-polymer and the sepiolite are stronger giving rise

to physical gels with a solid like behavior (G’ > G’’) and consequently, with lower

values of n’ and n’’. As for the system with the unmodified sepiolite, the

difference between n’ and n’’ diminish as the cross-linking reaction proceeds.

24

Figure 7. Variation of the power law exponents, n’ and n”, as a function of ε for

epoxy/D230/ B5 cured at 75 ºC in parts a) and b), and epoxy/MDEA/ B5 cured

at 130 ºC in parts c) and d).

Nevertheless at a certain reaction time, the difference between n’ and n’’ begins

to increase and the critical exponents do not crossover. The criterion of Winter

and Chambon is not fulfilled and the gelation point cannot be determined. To

study the effect of the filler concentration on the gelation time a new criterion

has to be established: the gelation time is considered as the time for which the

difference between n’ and n’’ reaches a minimum. The values of the gelation

times are shown in Table 2.

0.0

0.5

1.0

1.5

2.0

n´ 0.0 phr 2.5 phr 5.0 phr 7.5 phr

a)

0.0

0.5

1.0

1.5

2.0

n´´

J230 J230 B5 2,5 pcr J230 B5 5,0 pcr J230 B5 7,5 pcr

b)

-0.2 -0.1 0.0 0.1 0.20.0

0.5

1.0

1.5

2.0

n´

ε

0.0 phr 2.5 phr5.0 phr7.5 phr

c)

-0.2 -0.1 0.0 0.1 0.20.0

0.5

1.0

1.5

2.0

n´´

ε

0.0 phr 2.5 phr 5.0 phr 7.5 phr

d)

25

The data in Table 2 show that the addition of sepiolite B5 provokes a decrease

in the gelation times of the same order of magnitude as the addition of

unmodified sepiolite. As the Winter and Chambon criterion is not fullfiled, it is

not possible to calculate the value of the gel strength for the cured systems.

Neverthless, the lower values of the critical exponents obtained for the systems

with the modified sepiolite indicate that the gels formed with this clay are more

rigid.

Conclusions

This study shows the utility of rheological studies in understanding the effect of

nanoparticles on the gelation process in thermoset polymer nanocomposites.

The unfilled and the non-modified sepiolite epoxy systems obey the Winter-

Chambon criterion at the gel point since the loss tangents intersect at a single

point in multifrequency rheological measurements. However, for the organically

modified sepiolite a physical gel is formed and G’ is higher than G” for

concentrations higher than 5 phr when curing with MDEA and higher than 7.5

phr in presence of D230. The three parameters tanδ, n and Sg at the gel point

evaluate the gel properties. A more elastic gel with higher gel strength is

obtained in presence of the non-modified sepiolite as curing proceed in

presence of the aliphatic diamine, D230.

References

1 Pascault JP, Sautereau H, Verdu J and Williams RJJ, Thermoseting

Polymers. Marcel Dekker, New York (2002).

26

2 “Epoxy polymers: new materials and Technology” JP Pascault, RJJ

Williams Eds, Wiley-WCH, Weinheim, 2010.

3 Pinnavaia TJ and Beall GW, Polymer-Clay Nanocomposites. John Wiley,

Chichester (2001).

4 (a) Darder M, López-Blanco M, Aranda P, Aznar AJ, Bravo J and Ruiz-

Hitzky E, Chem Mater 18 (6): 1602-1610 (2006). (b) Kavas T, Sabah E

and Çelik MS, Cem Concr Res 34 (11): 2135-2139 (2004).

5 (a) Preisinger A, Clays Clay Miner. 6: 61-67 (1959). (b) Santarén J, Sanz

J, Ruiz-Hitzky E, Clay Miner. 38: 63-68 (1990).

6 Torro-Palau A, Fernández García JC, Orgiles-Barcelo C, Pastor Blas M

and Martin-Martínez JM, J Adhes Adhesives 17: 111 (1997).

7 Ruíz R, del Moral JC, Pesquera C, Benito T and González F,

Thermochim Acta 279: 103 (1996).

8 Bokobza L, J Appl Polym Sci 93(5): 2095-2104 (2004).

9 Bokobza L, Burr A, Garnaud G, Perrin MY and Pagnotta S, Polym Int

53(8): 1060-1065.

10 Bokobza L, Leroy E and Lalanne V Eur Polym J 45: 996-1001 (2009).

11 Roy N and Bhowmick AK, Polymer 51: 5172-5185 (2010).

12 Bilotti E, Zhang R, Deng H, Quero F, Fisher HR and Peijs T, Compos Sci

Technol 69: 2587-2595 (2009).

Con formato: Color de fuente:Automático, Español (España -alfab. internacional)

27

13 Rong J, Jing Z, Hong X and Zhang W, “Polyolefin-clay nanocomposites

and process for the preparation thereof. U.S. Patent 6444742, 2002.

14 García-López D, Fernández JF, Merino JC, Santarén J and Pastor JM,

Comp Sci Tech 70: 1429-1436 (2010).

15 Xie S, Zhang S, Wang F, Yang M, Séguéla R, Lefebvre JM, Comp Sci

Tech 67: 2334-2341 (2007).

16 Nohales A, Muñoz-Espi R, Félix P and Gómez CM, J Appl Polym Sci

119: 539-547 (2011).

17 Franchini E, Galy J and Gérard JF, J Colloid Interface Sci 329: 38-47

(2009).

18 Sangerano M, Pallaro E, Roppolo I and Rizza G, J Mater Sci 44: 3165-

3171 (2009).

19 Zheng Y and Y Zheng Y, J Appl Polym Sci 99: 2163-2166 (2006).

20 Winter HH, Polym Eng Sci 27: 1698-1702 (1987).

21 Hu X, Fan J and Yue CY, J Appl Polym Sci 80: 2437-2446 (2001).

22 Tung C-Y M and Dynes PJ, J Appl Polym Sci 27: 569-574 (1982).

23 Winter HH and Chambon F, J Rheol 30: 367-382 (1986).

24 Chiou BS, Raghavan SR and Khan SA, Macromolecules 34: 4526-4533

(2001).

25 Helmy AK and de Bussetti SG, Appl Surf Sci 255: 2920-2924 (2008).

28

26 Tung CYM and Dynes PJ, J Appl Polym Sci 27: 569-574 (1982).

27 Chambon F, Petrovic ZS, MacKnight WJ and Winter HH,

Macromolecules 19: 2146-2149 (1986).

28 Chambon F and Winter HH, J Rheol 31: 683-697 (1987).

29 Martin JE, Adolf D and Wilcoxon JP, Phys Rev Lett 61: 2620-2623

(1988).

30 Muthukumar M, Macromolecules 22: 4656-4658 (1989).

31 te Nijenhuis K, Advances in Polymer Science 130: 1-267 (1997).

32 Huang ML, Williams JG, Macromolecules 27: 7423-7428 (1994).

33 Girard-Reydet E, Riccardi CC, Sautereau H, Pascault JP,

Macromolecules 28: 7599-7607 (1995).

34 Lairez D, Adam M, Emery JR and Durand D, Macromolecules 25: 286-

289 (1992).

35 Becker O, Simon GP, Varley RJ and Halley PJ, Polym Eng Sci 43: 850

(2003).

36 Pethrick RA, Miller C and Rhoney I, Polym Int 59: 236-241 (2010).

37 Asif AA, John B, Rao VL and Ninan KN, Polym Int 59: 986-997 (2010).

38 Lange J, Johansson M, Kelly CT and Halley PJ, Polymer 40: 5699-5707

(1999).

39 Meng J, Hu X, Boey FYC and Li L, Polymer 46: 2766-2776 (2005).

29

40 Scanlan JC and Winter HH, Macromolecules 24: 47-54 (1991).

41 Kornmann X, Lindberg H and Berglund LA, Polymer 42: 4493-4499

(2001).

30

Table 1. Gel times at different criteria, critical exponent, n, tanδ, fractal

dimension, df, and gel strength, Sg for epoxy-sepiolite S9 cured with D230 at 75

ºC and with MDEA at 130 ºC.

Hardener D-230 MDEA

phr 0.0 2.5 5.0 7.5 0.0 2.5 5.0 7.5

tgel (min)

(G’=G’’, 5 rad/s)

54.2 51.1 46.7 42.0 102.0 91.0 80.3 75.2

tgel (min)

tan δ≠f(ω)

52.6 49.7 45.1 40.4 101.1 89.8 79.2 74.2

tgel (min)

n’=n’’

52.3 49.8 45.0 39.7 101.0 89.8 79.3 74.2

n 0.659 0.637 0.636 0.628 0.628 0.639 0.645 0.648

tan δ 1.685 1.559 1.554 1.512 1.512 1.570 1.603 1.620

df 1.80 1.83 1.83 1.84 1.84 1.82 1.81 1.81

Sg 522 1115 1556 1358 360 267 43.7 32.4

31

Table 2. Gel times for epoxy-sepiolite B5 cured with D230 at 75 ºC and with

MDEA at 130 ºC.

Hardener D-230 MDEA

phr 0.0 2.5 5.0 7.5 0.0 2.5 5.0 7.5

tgel (min)

(G’=G’’, 5 rad/s)

54.2 52.4 44.1 G’>G’’ 102.0 97.0 G’>G’’ G’>G’’

tgel (min)

(n’-n’’)minimun

52.3 49.7 42.22 33.3 101.1 98.0 96.1 82.3

Documents

Rheological study of the gel phenomena during epoxide